Peak expiratory flow (PEF) presents a circadian rhythm with a maximum in the afternoon, and a significant variability in its diurnal variations has been reported in normal subjects and in chronic obstructive pulmonary disease (COPD). In order to investigate whether passive smoking, active tobacco smoking, COPD and interstitial lung disease (ILD) are associated with changes in the circadian rhythm of PEF, five groups of adult male subjects, comparable for age, weight and height, were studied: group A: 30 clinically healthy subjects who never smoked, group B: 30 subjects passively exposed to tobacco smoking, group C 30 heavy smokers (> 20 cigarettes daily for at least 5 years), group D: 30 patients with nonasthmatic COPD (emphysema and/or chronic bronchitis), and group E: 15 patients with ILD (pneumoconiosis). Active tobacco smoking and exposure to passive smoking were assessed by the determination of the urinary cotinine concentration. A portable spirometer was used to measure PEF over a whole day, at 0.00, 6.00, 8.00, 10.00, 12.00, 14.00, 16.00, 18.00, 20.00, 22.00, and 24.00 h, all subjects leading a normal life. The 'mean cosinor' method was used for statistical analyses; the PEF variability was evaluated by the amplitude percent mesor (daily mean). All groups showed diurnal fluctuations in PEF values with significant (p < 0.05) circadian rhythms. The peaks of PEF rhythms occurred in the early afternoon, without significant (p > 0.05) differences between the groups. The cosinor mean was significantly (p < 0.05) lower in heavy smokers, in passive smokers, and in COPD patients than in controls. Controls, passive smokers, heavy smokers, COPD and ILD patients presented a PEF amplitude percent mesor (95% confidence limits) of 6.26% (range 4.57-7.95), 7.79% (range 5.07-10.51), 12.60% (range 7.61-17.59), 17.19% (range 10.18-23.50), and 3.98% (range 2.09-5.87), respectively, with significant differences (p < 0.05) between all groups, except between controls and passive smokers. These data suggest that tobacco smoke, both passive and active, does not modify the circadian peak of PEF, but modifies significantly its mesor and amplitude. In this respect, heavy smokers have the same pattern of COPD: lower mesor and greater amplitude; passive smokers present an intermediate situation. An increased diurnal variability in PEF could be considered as an early index of tobacco smoke damage and of developing COPD. When studying diurnal PEF variability, active and passive smoking habits should be considered.